The Agent Skills Directory

[DYNAMIC_EXECUTION]: The obsidian eval command allows the execution of arbitrary JavaScript code directly within the Obsidian application process. This provides full access to Obsidian's internal APIs and potential access to the underlying system depending on the application's configuration.
[DYNAMIC_EXECUTION]: The obsidian dev:cdp command provides access to the Chrome DevTools Protocol, allowing for low-level interaction with the application's rendering engine and UI components, which could be used for data extraction or UI manipulation.
[EXTERNAL_DOWNLOADS]: The skill includes commands such as obsidian plugin:install and obsidian theme:install, which facilitate the downloading and execution of third-party code from the Obsidian community registry.
[METADATA_POISONING]: The skill description and body repeatedly claim this is the 'official' Obsidian CLI. At the time of this analysis, Obsidian does not have an official CLI binary with these capabilities, making the claim deceptive and potentially leading users or agents to grant it higher trust than warranted.
[INDIRECT_PROMPT_INJECTION]: The skill is highly vulnerable to indirect prompt injection due to its core functionality.
Ingestion points: obsidian read, obsidian search, obsidian tasks, and obsidian template:read allow the agent to ingest untrusted data from the vault.
Boundary markers: The instructions lack delimiters or warnings to ignore instructions found within note content.
Capability inventory: Includes file system modification (create, append, delete), arbitrary code execution (eval), and system-level debugging (dev:cdp).
Sanitization: No evidence of sanitization or validation of the content read from the vault before it is processed by the agent.
[COMMAND_EXECUTION]: The skill facilitates the execution of the obsidian command-line utility with various parameters, including those that can overwrite files (--overwrite) or perform permanent deletions (--permanent).

obsidian-cli